https://www.mdu.se/

In this paper, we report our ongoing work on safe and secure controller synthesis for cyber-physical systems (CPS). Our approach separates the synthesis process into three phases, in which we alternatively perform exhaustive and selective exploration of the system’s state space. In this way, we combine the strengths of exhaustive search and learning to mitigate the state-space-explosion problem in controller synthesis while preserving the guarantee of safety and security. We implement the synthesis algorithms in the Rebeca (Reactive Objects Language) platform, which provides modelling, verification, and state-space visualization. We evaluate the new approach in an experiment, demonstrating the reduced number of explored states, which shows the potential of our approach for synthesizing safe and secure controllers for complex CPS.